| 摘要 |
PROBLEM TO BE SOLVED: To provide a KANA (Japanese syllabary)/KANJI (Chinese character) converter and a KANA/KANJI conversion method capable of further improving the correct answer rate of clause segmentation for segmenting a character string to clauses by using the input interval of character data and accelerating the processing speed of a KANA/KANJI conversion processing. SOLUTION: An input processing part judges whether or not inputted character data are a character for easily generating segmentation in front or at the back of the character such as a postpositional particle candidate character or the like (n7 and n8). When it is the character for easily generating the segmentation, the input interval of the character or the character inputted next to the character is extended from the actual input interval for α (n9). Then, the input interval is compared with a threshold value, and when the input interval is longer than the threshold value, the front of the character is turned to the candidate of the segmentation (n11-n13). |
